Home » Chinese military buys chips from AMD, NVIDIA and Intel for use in AI

Chinese military buys chips from AMD, NVIDIA and Intel for use in AI

China proved capable of producing its own chips, thus realizing CPUIt’s like him Longson 3A5000 That is GPUsIt’s like the JM9271which is equivalent to a GTX 1080 with a consumption of 180W. However, it seems that the chinese army interested in processors and graphics cards from the big companies that make these friesas amd, Nvidia and Intel. And it is estimated that they will be used for Artificial intelligence (IA).

In recent years, China has shown its ability to produce its own processors and graphics, which do not depend on foreign technology. So, by manufacturing their product locally, they could save huge costs and not have to go to companies like Intel or AMD. However, according to a recent report, the People’s Liberation Army (EPL), i.e. the armed forces of the Chinese country, have purchased several CPUs and GPUs from AMD, Intel and NVIDIA.

Chinese military prefers US chips over its own for use in AI

China United States AMD Intel NVIDIA AI Chips

This is stated by a research report from the Center for Security and Emerging Technology (CSET) from Georgetown University. And it is that, in what you can see buy files of the national army of the people of the republic of china. The report shows a total of 66,321 bids issued by the Chinese military since 2020, which are reduced to 21,088 search offers of chips and technology products.

After that it was searched for keywords such as CPU, GPU and others which resulted in 323 contracts of purchase. Of these, 24 mentioned products of the semiconductor market, such as graphics cards, SO Cyes fpgayes Finally, out of these 24, Eleven of them mentioned the number of units required by the Chinese army, with a total of 97 products. While that may seem too small a number to be of concern, CSET warns that the actual number of orders probably a lot elderly.

Indeed, the report itself only indicates that declassified files, it remains to add the classified records. Of these, it is estimated that from March to December 2020 alone, the Chinese military awarded nearly 2,000 classified contracts which could be chips for the AI.

Chinese military buys NVIDIA RTX 8000, Intel Xeon Platinum 8260 and AMD FPGA (Xilinx)

Chinese Army AI Chip Purchase Quantities in China

As seen in the shortened purchase list of the Chinese army, we have a total of 97 devices if we count the quantities indicated. Among these we are struck by those 4 Intel Xeon Platinum 8260 of 24 hearts Yes 48 threads a 2.4GHz base and 3.9GHz in fashion turbo. Although they have been on the market for a few years and are not the highest range, when they were launched in 2019 their official price was $5,030.

Another interesting product are those 10NVIDIA Titan Vannounced in 2017 with a price of $2,999 each. These have 12 GB by heart HBM2, 5120 CUDA cores Yes 640 tensioner cores. Already less expensive we find other NVIDIA products like 4 Jetson TX1at the price of $599. Unfortunately, we don’t know the number of GPUs TeslaV100 (almost $10,000) nor of the Quadro RTX 8000 ($5,500).

Regarding AMD, we only found interesting FPGA chips for the Chinese army, considering that Xilinx was purchased by AMD itself. Now, what is curious is that these 11 contracts managed to overcome the american security. And it is that the US government, through the Department of Commerce, monitors and punishments to any entity that collaborates with the Chinese armed forces.

In other words, the intermediaries who succeeded in supplying these chips to the Chinese army do not seem to have been sanctioned. One of them would be sitonholly (Tianjin) Co., Ltd., which is a partner Official distributor of NVIDIA and provides AI products to Chinese universities. The other, called Beijing Hengsheng Technology Co., Ltd., appears to be a frontsince his email address has been used by dozens of Beijing technology consulting firms.

About the author


Add Comment

Click here to post a comment

Your email address will not be published. Required fields are marked *